Reference

Top-Level Functions

The primary gql package includes everything you need to execute GraphQL requests, with the exception of the transports which are optional:

  • the gql method to parse a GraphQL query

  • the Client class as the entrypoint to execute requests and create sessions

class gql.Client(schema: Optional[Union[str, graphql.type.schema.GraphQLSchema]] = None, introspection=None, transport: Optional[Union[gql.transport.transport.Transport, gql.transport.async_transport.AsyncTransport]] = None, fetch_schema_from_transport: bool = False, execute_timeout: Optional[Union[int, float]] = 10, serialize_variables: bool = False, parse_results: bool = False)

Bases: object

The Client class is the main entrypoint to execute GraphQL requests on a GQL transport.

It can take sync or async transports as argument and can either execute and subscribe to requests itself with the execute and subscribe methods OR can be used to get a sync or async session depending on the transport type.

To connect to an async transport and get an async session, use async with client as session:

To connect to a sync transport and get a sync session, use with client as session:

__init__(schema: Optional[Union[str, graphql.type.schema.GraphQLSchema]] = None, introspection=None, transport: Optional[Union[gql.transport.transport.Transport, gql.transport.async_transport.AsyncTransport]] = None, fetch_schema_from_transport: bool = False, execute_timeout: Optional[Union[int, float]] = 10, serialize_variables: bool = False, parse_results: bool = False)

Initialize the client with the given parameters.

Parameters
  • schema – an optional GraphQL Schema for local validation See Schema validation

  • transport – The provided transport.

  • fetch_schema_from_transport – Boolean to indicate that if we want to fetch the schema from the transport using an introspection query

  • execute_timeout – The maximum time in seconds for the execution of a request before a TimeoutError is raised. Only used for async transports. Passing None results in waiting forever for a response.

  • serialize_variables – whether the variable values should be serialized. Used for custom scalars and/or enums. Default: False.

  • parse_results – Whether gql will try to parse the serialized output sent by the backend. Can be used to unserialize custom scalars or enums.

execute(document: graphql.language.ast.DocumentNode, *args, **kwargs)Dict

Execute the provided document AST against the remote server using the transport provided during init.

This function WILL BLOCK until the result is received from the server.

Either the transport is sync and we execute the query synchronously directly OR the transport is async and we execute the query in the asyncio loop (blocking here until answer).

This method will:

  • connect using the transport to get a session

  • execute the GraphQL request on the transport session

  • close the session and close the connection to the server

If you have multiple requests to send, it is better to get your own session and execute the requests in your session.

The extra arguments passed in the method will be passed to the transport execute method.

subscribe(document: graphql.language.ast.DocumentNode, *args, **kwargs)Generator[Dict, None, None]

Execute a GraphQL subscription with a python generator.

We need an async transport for this functionality.

gql.gql(request_string: str)graphql.language.ast.DocumentNode

Given a String containing a GraphQL request, parse it into a Document.

Parameters

request_string (str) – the GraphQL request as a String

Returns

a Document which can be later executed or subscribed by a Client, by an async session or by a sync session

Raises

GraphQLError – if a syntax error is encountered.